ysh329 OpenCL-101 issues

ysh329 / OpenCL-101

Learn OpenCL step by step.

131 stars 29 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Preferred / native vector sizes?

#19 ysh329 closed 7 years ago
1
Performance difference between two write methods in kernel

#18 ysh329 closed 7 years ago
2
OCL error: implicit declarations are not allowed

#17 ysh329 closed 7 years ago
2
Strange problem about index of vector variable

#16 ysh329 closed 7 years ago
1
Optimize matrix transpose

#15 ysh329 opened 7 years ago
1
Follow CPU GEMM optimization guide

#14 ysh329 closed 7 years ago
2
Implement Conv progress

#13 ysh329 opened 7 years ago
0
discover optimization strategy for no local memory situation

#12 ysh329 closed 7 years ago
2
benchmark for various type (floatN, intN, halfN, doubleN, shortN) using naive implementation

#11 ysh329 closed 7 years ago
6
Unstable benchmark result for GFLOPS

#10 ysh329 closed 7 years ago
1
benchmark.py: for auto-benchmark for different variable type (half, short, int, float, double) with matrix size (2048 * 2048)

#9 ysh329 closed 7 years ago
1
Define half16 as input-variable-type for kernel function

#8 ysh329 closed 6 years ago
1
[bandwidth] Bandwidth for typeN and compare with clpeak result

#7 ysh329 closed 7 years ago
2
Bug of floatN/halfN support: fail to copy data from device to host

#6 ysh329 closed 7 years ago
3
Performance loss caused by input type of kernel function

#5 ysh329 closed 6 years ago
2
make warning: ‘clCreateCommandQueue’ is deprecated [-Wdeprecated-declarations]

#4 ysh329 closed 7 years ago
3
double free or corruption (!prev) Aborted (core dumped)

#3 ysh329 closed 7 years ago
4
Support halfN precision for GPU and CPU

#2 ysh329 closed 7 years ago
1
Add other OCL mini-projects/demos from github

#1 ysh329 opened 7 years ago
4